Assessment and Comparison of Meta-features for Educational Chatbots Data and Survey Data

Abstract

Usage of chatbot platforms is acquiring great attention at all levels of education. Human-chatbot interactions generate huge amounts of data which are a valuable source of information, when properly analyzed by means of data and text mining. One of the most challenging tasks in the mining process is the selection of the appropriate algorithm for a data set at hand. This is a complex task and depends on characteristics of the dataset used in the analysis. Those characteristics are formalized through meta-features. In this paper, we identified meta-features of chatbot and survey data. As a case study, we evaluated two data sets and identified their general meta-features along with discussion. This is, to the best of our knowledge, the first examination of meta-features for chatbot interactions data and their comparison with survey data.